Young women are at disproportionate risk of STIs and their consequences, with the highest rates of chlamydia and gonorrhea, as well as outcomes such as pelvic inflammatory disease, infertility and cervical cancer. This disease burden has a disparate on some groups of young women, such as racial/ethnic minorities and women of low socio-economic status. To date, most research and prevention programs have focused on individual- level predictors of sexual risk behavior measured at a single time point (e.g., behavioral intentions, attitudes). However, individuals' intentions to engage in risky or protective behaviors may vary across contexts and over the course of relationships. By applying a cutting-edge analytic technique to richly detailed, intensive longitudinal data, we will extend research on sexual risk behavior to address the relationship dynamics that influence these behaviors. This project will make use of data collected weekly from nearly 1,000 young women over 2.5 years, including detailed information about their relationships, sexual behavior, contraceptive use, attitudes and intentions. We will apply a novel analytic technique, the time-varying effect model, to understand women's sexual risk behavior over the course of the relationship with a sexual partner. This project has three specific aims. First, we will model patterns of young women's non-use of condoms over the course of sexual partnerships. Second, we will examine how women's intentions to use condoms differentially predict condom non-use over the course of sexual partnerships. Finally, we will delve into sources of disparities in sexual health outcomes by examining how these patterns of sexual risk behavior may differ by race/ethnicity, socioeconomic factors and urbanicity. By modeling how sexual risk behavior evolves over the course of a relationship, the proposed project will inform the creation of more effective prevention programs that are targeted to high risk populations and relationships contexts.